Are Novel Junctions Only Give Novel Transcripts?

Question

[Ngsnewbie]

In a well annotated GTF/GFF file, The genomic coordinates of exons of genes are present. A gene can produce different isoforms by a mechanism known as alternative splicing. If we try make different permutation/combinations to join the exons of a gene in-silico, we can get different isoforms of a gene. Suppose we get 14 isoforms of one gene. Out of 14, 9 isoforms/transcripts are reported in ensembl/NCBI-genpept or other protein seq repository while 5 are not reported till date. so are they 5 transcripts/isoforms are novel which are being come out from known exon junctions of available GTF/GFF file?

or

In RNA-seq novel junction finding strategies, we came across with some novel junction (new splice sites in known exons) coordinates which are not specified in the GFF/GTF file. Some isoforms may be generated by splicing from these novel junctions. so my question is that are these transcripts will be considered novel or the above mentioned 5 transcripts or both?

Answer 1

[Damian Kao]

The 5 possible isoforms in the first scenario are just possible alternative splice forms. Without any experimental validation or RNA-seq data to back it up, they are not considered novel. For example, if there were split reads or pair-end reads to back up the 5 possible isoforms in the first scenario, then they would be considered novel.

The second scenario where new splice sites are discovered leading to evidence for new alternative splicing would be considered novel transcripts.

[Ngsnewbie]

Thanks DK, I got the answer. BTW , I use cufflinks class codes "j" for identification of novel elements. Is there any other way to pick them up? please put some light in this context.